Recent Posts

Sunday, January 31, 2016

Rise Of Robot Lettuce

A story about a robot-run lettuce farm that can produce 10 million heads a year has been making the rounds (h/t Mike Norman Economics). Although interesting, you cannot extrapolate this to imply that all farming jobs are at risk. The article is interesting in how the techno-enthusiasm covers up the interesting economics. (This article could be viewed as a post-script for my review of the book Rise of the Robots.)

Summary Of What Is Happening

I am relying on the information in the article written by Adele Peters; the risk is that I am misinterpreting what she wrote.
  • The factory will be produce 30,000 heads of lettuce a day, or 10 million a year; it is scheduled to open in 2017.
  • Spread, the Japanese company that is building the factory, already runs a highly automated indoor lettuce growing operation in Japan. The new factory is projected to save 50% on labour costs (relative to that plant). If compared to a traditional lettuce farm, the labour saving is presumably much higher. In other words, it is a good bet that most of the potential job losses would have already happened.
  • The plant covers less acreage than a soil farm, since the lettuce is grown in a vertical stack. However, this implies that artificial light must be used at all times for plant growth. (A traditional greenhouse operation is not vertical, and will supplement sunlight with artificial light only when necessary.)
  • The environment is largely sealed, reducing the risk of contamination, disease, and pests. (Lettuce is highly vulnerable to contamination since it is typically eaten raw.)

Robots: A Red Herring

The appearance of robots here is largely a red herring. Vertical farming has been around for awhile, and it can be viewed as a means of converting concentrated energy into food. Even if the lighting is "ultra-efficient," electricity still has to be supplied. Traditional farming is still the most efficient way of utilising solar energy: you let the plants do the work.

Based on the article, their existing factory was just competitive with other lettuce producers. Indoor farming uses less acreage and labour, but it is capital-intensive. That is, there is a lot of labour content that is embedded in the extra capital goods that are required. Those capital goods are relatively cheap so long as resource limitations do not bite. Doubling the cost of electricity is going to destroy the profitability of the factory, whereas that would have no effect on the cost of production for a traditional farm.

However, the extensibility of this factory to other crops is debatable. It is hard to visualise grains or corn being grown under artificial lighting unless we have access to cheap fusion power. Grains provide the bulk of the food supply, either directly or indirectly. The only jobs at immediate risk in the agriculture sector with technology very similar to this are those people who are solely employed in harvesting indoor plants. (Admittedly, there are other technologies for outdoor harvesting.)

The article did not specify this, but the pictures were of leaf lettuce, and not iceberg lettuce. Leaf lettuce is just uprooted and put into a bag; iceberg lettuce has covering leaves removed before being sent to consumers. It is unclear whether robotic arms have the dexterity to do such operations yet. Other crops require dexterity when harvesting, and figuring out to harvest them is an ongoing challenge. (The book Rise of the Robots discusses this; roboticists have figured out work arounds for some crops, such as shaking the trunk of a tree to drop the crop.)

Inorganic Farming?

The article states that this is "sustainable" farming, which is a sign that this has become a meaningless term. Indoor farms can recycle water that is not absorbed by plants, unlike the lettuce farms in the deserts of California that require extensive irrigation. However, being more sustainable than mono-cropping leafy vegetables in a desert is setting the bar very low. The system requires concentrated energy for plant growth, factory operation, and to manufacture the modern engineering materials used by the system. Furthermore, the manufacturing supply chain is dependent upon a multitude of inputs, each facing its own depletion curve. Although the technology makes economic sense now, I doubt that robot-grown lettuce will feature highly in diets 150 years in the future.

The term "organic farming" has become meaningless due to fights over politics and marketing rights. However, this is an attempt at what might be described as inorganic farming -- all inputs into the plant are biologically sterile. Given the food safety concerns around lettuce, this is not a bad idea. A large mono-crop of indoor lettuce is vulnerable to blights like fungus, and even cosmetic damage eliminates the crop's consumer acceptability. Very simply, they cannot just dump compost on the plants and expect anything other than a disaster. The need for sterile inputs creates costs and labour inputs that are not directly visible.

On the whole, industrial farming is a race between improved automation and genetic modification versus resource depletion. Even if technology is likely to win over the next few decades, the mathematics of depletion will make it progressively harder to stay ahead. I have few worries about job destruction in agriculture; the real worry is that it might start creating too many jobs.

For those who are interested in the topic, I would recommend the book The End of Food, by Paul Roberts discusses the trends shaping the food industry. Despite the rather alarming sounding title, it is a balanced account of the risks and opportunities in the food industry. It was published in 2008, and so presumably there have been some advances since then, but it is unlikely that the outlook has markedly shifted. My feeling is that "Peak Food" may be visible even before energy constraints, and any rollback in agricultural productivity would have quite large economic consequences.  

See Also:

(c) Brian Romanchuk 2015


  1. Problem with 'Peak Food' is if it's coupled with 'Peak Fresh Water' (which may happen even before). This could pose a very apocalyptic scenario in certain parts of the planet with all the social (and ecological) problems one would expect in that case (massive migrations that will dwarf the current ones in the Middle East, social collapse, famine and malnutrition, increasing diseases, exploding natality -because that's what organisms do when the environmental pressure increases to maximize the chance of survival/reproduction- and thereby mortality, etc.).

    The distance part of the global population have to walk to get their daily rations of water is increasing right now. This is like having to revert to labor intensive agriculture: it's time (and energy) consuming, and is a low productivity task, meaning economic regression.

    1. I tend to lump "Peak Water" with "Peak Food" (which is why I use the label "Peak Everything"); we hit "Peak Food" as a result of any number of constraints on inputs. As people keep re-discovering, the global trade in grains is disguised trade in water. However, I'd agree that in a disaster situation, a loss of water becomes dangerous far more rapidly than a breakdown in the food supply.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.